Intra-prediction Mode Determination for Image Coding Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image prediction methods, such as H.264, require long code lengths for encoding mode information, leading to increased code amounts and poor encoding efficiency due to inaccurate reference mode information, especially when adjacent blocks are encoded using inter-frame prediction, and the method of determining prediction methods for adjacent regions is not always suitable for the target block.

Innovation Solution

The proposed solution involves an image prediction encoding device and method that generates intra-frame prediction signals by determining an adjacent region prediction method using a proximate pixel group and predicting the target region prediction method based on this, reducing the code amount for mode information identification and improving prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple prediction methods are provided to enhance prediction accuracy, then prediction precision is improved, but code length for mode information increases

Engineering Contradiction:
Improveprediction accuracyVSAvoidcode amount
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent applies preliminary action by determining the prediction method for the adjacent region before determining the prediction method for the target region. This allows the encoder to use the already-determined adjacent region prediction method as reference information when encoding the target block, reducing the number of bits needed to signal the target block's prediction method while maintaining accurate prediction.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If reference mode information from adjacent blocks is used, then encoding efficiency is improved, but accuracy deteriorates when adjacent blocks use inter-frame prediction

Engineering Contradiction:
Improveencoding efficiencyVSAvoidreference mode information accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent introduces an intermediary approach by determining the adjacent region prediction method independently first, using it as a reference, but allowing the target block to select its own optimal prediction method from multiple candidates. This intermediary step provides a baseline reference while maintaining the flexibility to achieve accurate prediction for the target block, even when adjacent blocks use inter-frame prediction.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Ease of manufacture

If prediction method is determined using pixel group not in contact with target block, then processing is simplified, but prediction accuracy decreases

Engineering Contradiction:
Improveprocessing simplicityVSAvoidprediction accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The patent segments the prediction process into two distinct parts: first determining the prediction method for the adjacent region using its own pixel groups, then using that result as reference for determining the target block's prediction method. This segmentation allows each region to be processed with appropriate accuracy while maintaining overall system efficiency.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP2503781B1Devices, methods, and programs for INTRA-prediction in image coding
Publication Date: 2017.03.08 NTT DOCOMO INC
  • EP2503781B1 patent drawingFigure 1
  • EP2503781B1 patent drawingFigure 2
  • EP2503781B1 patent drawingFigure 3

AI summary

An object is to efficiently depress mode information for identification of a prediction method even if there are a lot of prediction methods provided as to methods of generating intra-frame prediction signals in a pixel domain. An image prediction encoding device 10 has a block dividing unit 102, an intra-frame prediction signal generation method determining unit 105, an intra-frame prediction signal generating unit 106, a subtractor 108, a transform unit 109, a quantization unit 110, and an entropy encoding unit 115; the intra-frame prediction signal generation method determining unit 105 derives an adjacent region prediction method for generation of an intra-frame prediction signal highly correlated with pixel signals of an adjacent region, using a proximate pixel group to a target region, out of a plurality of first prediction methods, and predicts a target region prediction method for target pixel signals, based on the derived adjacent region prediction method; the intra-frame prediction signal generating unit 106 generates an intra-frame prediction signal for the target region, based on the target region prediction method.